THE BOOK--Playing The Percentages In Baseball

Agree. If you take Fangraphs WAR calculations, and replace Batting Wins with WPA, you get Gutierrez head and shoulders above everyone (largely due to a great UZR, and high Clutch factor). Figgins is next, who I think personally is a better choice for a Mauer MVP argument then say, Tex.

What about Felix? Isn’t he close to Grienke?

Felix: 5.9 WAR
Grienke: 8.3 WAR

That’s an easy one.  In about the same innings, Greinke has allowed 11 fewer runs despite pitching in a worse park for pitchers, and getting vastly inferior defensive support.

"(I think I even invented it, but maybe one of my readers did, and I modified it?  I don’t remember any more.)”

I don’t recall which discussion led to the current form of Clutch [WPA/pLI - WPA/LI], but I’m almost positive that the immediately prior version [WPA - WPA/LI] and certainly the two before that [(WPA - OPSWins*LI) and the original (WPA - OPSWins)] originated in posts here by Tango.

You know what though?  I’m fine with making the MVP position player/hitters only, and i think most of us are too.

The reason being that while there are hitter only awards (Sliver Sluggers, Hank Aaron Award), they don’t have the prominence of the pitcher version, the Cy Young.  So why not just keep things as they functionally are where the MVP is the most valuable position player and the Cy is the Most valuable Pitcher. 

The Hank Aaron doesn’t necessarily mean the most valuable position player (no fielding, no accounting for position), so there’s not really any overlap and i think it’s fine.

Why do we care if he performs in the clutch?  Is this just because clutch is in the traditional MVP definition?

Why do we care if he performs in the clutch?

Because performing well in the clutch helps win ballgames more than not performing in the clutch.

I think the defensive stats have grossly undervalued fielding at 1B.  I look at the Yankee defense since adding Teixeira, and fielders no longer have to worry about making the perfect throw, errors are down and more plays are being made, and runs are being saved but not counted.

As a Red Sox fan I felt Youk was MVP last year, guess he is not cuter than Dustin though, and he has missed too many games in 2009 to be considered. Still, he is the Red Sox MVP since he allows them to rest Lowell and give playing time to V-Mart at 1B so he does not burn out his bat catching every game, plus he has been the teams most consistent hitter, H-A, LHP -RHP, 1st half-2nd half.

Joe Mauer presents a similar issue in terms of playing time, great hitter but he has missed 25 games.  Teixeira has played every game except 5.  Of course, he has too many RBI’s so he probably won’t even be considered by those who have swallowed the myth that RBI’s mean nothing.

Also, with the unbalanced schedule, lets use the same argument being used against Greinke, which is quality of opponents.  Mauer has only 38 AB (BA 263) against the top 10 starters in the AL (based on ERA and according to B-Ref pitching matchups). He does not seem to have faced Greinke at all.  Doesn’t mean much, but its interesting.

Pedroia lowered the bar with his MVP award, so it would not surprise me if Jeter gets it, and the gold glove too (defensively he is much better, maybe Teixeira relaxes him more than Giambi).  I pay no attention to these awards anymore so I could care less who wins until they upgrade the voters.

I don’t understand the use of clutch either.  I thought with clutch its a very small repeatable skill, if a skill at all.  Why would we reward players for essentially being lucky in clutch situations?

The award is Most Valuable for this particular season. It is not Best Player. I dont think anyone here would use one season worth of data to determine the Best Player in a particular league. As a result, I think we have to expect that the MVP award will have a certain amount of ‘luck’ attached to it. If a ‘clutchy’ stat like WPA measures wins perfectly, it measures value - regardless of how much luck is involved.

Willie Bloomquist is the quintessential replacement-level player.

If he manages to get a .400 OBP with a .500 SLG, we know that a huge part of that was pure luck.  Do we reward him… or not?

If someone hits 40 HR, and all of them occur with at least a runner on, and all of them occur with his team tied, or down by 1 or 2, and all of them occur in the 7th or later innings, do we reward him… or not?

The question is: are we rewarding the best player (in which case, we just give the MVP to Pujols before the season even starts), or do we reward the player involved in the best performance, regardless of how much his skill contributed, and we simply presume that because he was involved, that he gets the lion’s share credit for it?

And if we do that, if we simply reward performance, and attribute that performance to the player (rather than luck), then how far do we go?  Do we go all the way, and count 40 HR with men on base more than we’d count 40 solo shots in blowouts?

The point of this exercise is to establish BEFORE-THE-FACT the objective and constraints.  Once you’ve set up the process, however you as an individual would like, then you simply follow the process to its conclusion.  That is, you cannot change the process after you see the results.

If your process says that Chone Figgins is the MVP, then he’s the MVP, even if it doesn’t make “sense”.  If it says that it’s Mauer, or Jeter, or Teix, then that’s what it is.

But, the one thing you cannot do is start with the player, and then construct the argument for or against him.  You won’t be consistent if you do that.  Not unless you lay out your process first.

So, lay out your process, and then follow the evidence.

it seems arbitrary to me to assign greater value to performance that takes place late in games.  games (or seasons, or careers, for that matter) should be viewed as an aggregate, not individual pieces.  the perceived value of a late inning home run is dependent on the score, which is a result of all the previous innings.  decreasing mauer’s value based on not being clutch, comes off, in my opinion, as a use of data manipulation to lessen his value. the missed time and his use as a dh, however, are valid conversations.  that being said, i think the mvp award should be shared by dollars and new yankee stadium.

I’d like to know more about your Clutch definition before painting Mauer with this brush and automatically assuming that he is not “clutch.”

One thing I don’t like about the clutch/not-clutch argument is that it assumes that the AB where Joe Mauer struck out in the 8th inning of a game the Twins lost 6-5 is more valuable than the AB in the second inning of a game the Twins won 6-5. That HR by Mauer in the game the Twins won contributed just as much to the Twins win as the RBI in the 8th by Nick Punto that won the Twins the game. Saying that Punto’s hit had some kind of “clutch-y” feel to it gives Punto credit for having some kind of extra magical ability that I’m certain Nick Punto doesn’t have.

I don’t know if the word “clutch” is scaring some of the commenters, but Mauer is tied for 5th in the AL in WPA (Bay, Teixeira, Damon, A-Rod, t-Abreu).

You may not like that WPA places greater value on later innings, but the fact remains that a hit in a late-and-close situation does in fact have a greater effect on a team’s chances of winning the game than a hit in an early-and-close situation.

Or perhaps you can think of it as a rough measurement of the perceived “pressure” of a given at-bat.  That may or may not carry any weight for you.  We’re not measuring ability to perform, just actual performance.

And no one is giving any extra credit for clutch ability, just for the actual results of their actual at-bats.  If Mauer smokes a line-drive that’s caught for an out in a tie game, but Damon gets on base on a broken-bat dribbler to third, which AB was more valuable to his team?  But no one would say Damon is a better hitter.

Smith/13:

games (or seasons, or careers, for that matter) should be viewed as an aggregate, not individual pieces.

Untrue.  This is a matter of opinion, not fact.  Therefore, there is no “should”.  Rather than doing like a politician, and saying what someone else can’t do, why don’t you lay out your own process as to what you do want to consider.  You lay it out first, and then, follow the evidence.

Mike/14:

One thing I don’t like about the clutch/not-clutch argument is that it assumes that the AB where Joe Mauer struck out in the 8th inning of a game the Twins lost 6-5 is more valuable than the AB in the second inning of a game the Twins won 6-5.

For the very reason that you do NOT like it, that is the reasoning one uses to like it!  Again, this is a position of opinion, not fact.  You don’t have to like it.  You create your process as to what you want to consider, and then go from there.

There is really no reason whatsoever to take issue with how someone else handles a component, as long as they are consistent with it. 

What you can take issue with is if someone starts with the answer, and then constructs the “right” questions.  That is b.s.

Punto credit for having some kind of extra magical ability that I’m certain Nick Punto doesn’t have.

Again, he doesn’t have to have an ability.  Suppose I win the lottery.  Should I give the money back because it was pure luck?  Or, can I start dating starlets (wife’s objections notwithstanding)?

One can decide to create his process that does not reward ability or skill, but performance and results, regardless of how much of an influence the player had.

If Mauer smokes a line-drive that’s caught for an out in a tie game, but Damon gets on base on a broken-bat dribbler to third, which AB was more valuable to his team?  But no one would say Damon is a better hitter.

Under a certain framework, you can give Mauer a plus for doing that, you give the pitcher a minus, and you give the fielder a plus to cancel it all out so that you are left with an ordinary out.  For example, if the end result of that play is -.05 wins, then you might go:

+.10 Mauer
+.05 pitcher (from Twins’ perspective)
-.20 fielder (from Twins’ perspective)
------
-.05 from Twins’ perspective

I’m not saying that you have to do this.  I’m saying this is one valid process to do it.

I suppose this is basically just about how one defines “MVP”......everyone seems to have different interpretations on exactly what that means.....and no one is right or wrong.

There is really no reason whatsoever to take issue with how someone else handles a component, as long as they are consistent with it. 

What you can take issue with is if someone starts with the answer, and then constructs the “right” questions.  That is b.s.

Agreed.  And sadly I think this is what happens all too often.  Most of the time, people start with the player, not the process.

It’s not even that everyone has their own definition, but it seems that everyone doesn’t like everyone else’s definition, without even bothering to define their own!

It’s the sports-equivalent of pornography.

The difference between art and porno is that art has purpose, thought, depth, and creativity, while porno is just a means to an end.

Rosenthal and his ilk give us porno.  And then we are shocked it’s not art!

I’ve made the argument on Baseball Musings, unfortunately the award is called “Most Valuable Player”.  Until they change the name, that means trying to give the award to the player who adds the most value to a team. 

The award has been given to players on last place teams.  I would argue that if a team finishes last, no player on that team is valuable.  Without that excellent performance, the team would still finish last.  Now that eight teams are in the playoffs, it makes sense to restrict the award to players on the playoff teams, pre 1994 you can consider players on teams in division races in September, but under the pre 1994 divisions the Twins would be no way near division contention.

In a just and fair world, pitchers would win the “most valuable player” award 40% of the time, but we all know they have their own award, so its given to pitchers occassionally, seemingly at random.

The Yankees are going from 89 wins to 94 wins plus this year, there are lots of reasons for this, but everyone agrees replacing Teixeira is a key reason (remember Sabathia essentially replaced Mussina, not really an upgrade considering Moose’s last season.  The Yankees improved their bullpen and bench and got better performances from their second tier players).  Abreu has not gotten sufficient credit for his role with the Angels.

I would prefer instituting two Cy Young type awards for position players, one for the infield and one for the outfield and DH. They would include defense as well as hitting.  But what we have is the MVP, and I don’t think we should pretend it is really the award it should be.

I have to agree with Tango /12. Luck plays a great role in baseball and shouldn’t be discounted when looking at achievements. However, it should be including in discussions about future performance, because we clearly can’t count on that luck again.

“Clutch” shouldn’t be counted for duplication in future performace, but it should count heavily in our evaluation of past performace.

@ Tangotiger:
What you can take issue with is if someone
starts with the answer, and then constructs
the “right” questions.  That is b.s.

But isn’t this the whole point of journalism, to construct a narrative after the fact?  If you have journalists give this award, that is precisely what they are going to do.

Right or wrong, a lot of people--pretty much every baseball fan outside of Boston and New York--follow baseball for the human interest elements, not to see if their team will win the World Series that year.

(I do think that Mauer is the most valuable player, though.)

I have no problem if Rosenthal wants to create a narrative, write fiction, or whatnot.

What I do have a problem with is if he says that it’s my fault if I think his story s-cks.  Next thing you know, he’ll be shouting “You lie!” at me, and blame saberists for creating the environment that forced him to say it.

And yes, the MVP award is nothing more than a way for the BBWAA writers to get 5 free articles written and talked about ad nauseum, each and every year.  Really, you just have to fill-in the blanks, as the story’s already been written.

This is why I find even the discussion of it absurd.  It’s like we’re sheep to these writers.

Using one season clutch makes sense, because that’s actually how a player has performed in a clutch, minus BABIP luck.  One season UZR, on the other hand, is a little different.  We don’t know that Ben Zobrist is 23 runs above average in the field.  We have a limited sample and need to regress that number. I also find it hard to believe that Derek Jeter is now 5 runs above average in the field.  His number should be regressed back to his career norm.

Also, and I know people will disagree with me on this, but I think that Greinke’s HR/FB rates should be regressed some.  I know the arguements, MVP is how you performed this year, not how you’re projected to perfrom.  I just believe that his low HR/FB is a product of luck, and if he performed the exact same over the course of another season, would go up.

Why not regress Mauer’s OBP then?

I think that OBP is perceived as more under a hitters control that HR/FB, which is a “luck” stat in most cases.  Of course, OBP is influenced by BABIP, so it’s a little muddled. 

In any case, I wouldn’t regress anything, expect to try and take defense out of the equation for pitchers without neutralizing timing.

I think post #25 helps me articulate why I don’t see it as valid to use clutch in any system.  And is seems to me there is one correct system for however you decide to define “valuable player”.  Its the definition that causes there to be more then one system possible.

To use clutch in the definition of “valuable” implies that you actually do believe that clutch hitting exists.  Because the situation the individual performer came up to bat or pitch in was not in their control.  You can’t give them credit for being allowed to bat in high clutch situations.  It is almost completely a product of other people’s performances that produces clutch situations.

So that leaves us with two options for defining “valuable”.  Straight performance, how much each event was worth regardless of situation.  Or performance minus luck.

If you choose straight performance you simply add up the events and, using your favorite holistic Win total, give the best man hit due.

But, if we are going to introduce the fact that we don’t want luck in our system, than we have to debate how much of the seasons performance to we really believe is luck related.  How much we are going to allow to effect “valuable”.  When post #25 says “his low HR/FB is a product of luck” he is beginning to create his parameters for that.

I would argue that you should not allow luck in either if you are determining a particular season’s value.  As someone pointed out, if you remove luck then you are left with “best player”.  And if that is the case we do as someone suggested and give the award to Pujols before the season starts.  Or in the rare case that someone gets better than him by seasons end in terms of true talent, we give it to them.

Since we give the award at the end, I think that implies luck IS allowed to play a factor.  And so I believe adding up the total value of a player’s contributions independent of situation is the best system for determining the MVP.  And also a very boring one.

So alternatively I’d propose a system that disallows all calculations from MVP discussions so that there can actually be a discussion.  You can use totals, but not calculations.  Mauer for his game calling says I!

#25/Steven, I don’t necessarily agree that a player’s clutch performance is an accurate portrayal of what happened (aside from BABIP luck) and can be taken at face value while UZR (or whatever defensive ratings you want to use) is not and needs to be regressed.  A player’s UZR is going to have sample error, and to get an estimate of a player’s talent level, you have to regress.  There is also some uncertainty, as you say, to whether the player was really that many runs above or below average because of the measurement error.  But that issue also presents itself in measuring a hitter’s clutch performance.

“Clutch” measures what actually happened, true.  So does UZR.  Either a fielder made a play or he didn’t.  UZR gets that down pretty well.  Where the measurement error comes in and the uncertainty about the value comes from is in trying to decide how much each of those plays is worth, relative to the average fielder.  We know that not everyone receives the same batted ball distribution, so what we really want to ask is not whether a fielder made more plays than average or not, but, given the exact same distribution of opportunities, whether the fielder made more plays than an average fielder would have.  There is some error in how we measure the difficulty of each opportunity and the value relative to average.

Hitting stats take out the measurement error of assigning difficulty by simply ignoring that step altogether.  They just assume that everyone’s hitting opportunities over the course of a season are roughly the same, which is a much safer assumption than with fielders.  They just ask who did what more than average, distribution of opportunities aside.  If you really want to know how someone performed relative to average, though, you are still concerned with how someone hit not just relative to average, but relative to how an average hitter would perform given the exact same distribution of opportunities.  This introduces a problem with looking at a player’s clutch ratings:  what if the distribution of pitchers or pitches he faced was such that his opportunities in high leverage situations were simply more difficult than normal?  There is no way to tell whether or not that is the case from just looking at his clutch rating.

So say Derek Jeter rates at +5 in the field because we overestimated the difficulty of his batted ball distribution.  This means that our stat is wrong in how Jeter compares to average, because the average player would have actually done better given Jeter’s distribution than we thought.  Now say that Joe Mauer rates poorly in the clutch because he has faced more tough pitchers or even just more nasty pitches in high leverage opportunities than in low leverage opportunities.  Obviously, we would expect any hitter to do worse in his high leverage opportunities given that kind of distribution.  So our stat says he did worse than average in Clutch performance because we assume his distribution is average, but if it isn’t, our stat is going to be wrong in how it rates him relative to average.  He did worse in Clutch situations because anyone would have, which means he might actually be average in Clutch, not a win below average or whatever his rating is.

So there is still going to be measurement error in clutch ratings, just like in UZR.  Both measure what actually happened, and both have a measurement error in translating what actually happened to rating players relative to average.  So if your process tells you that you can’t fully trust UZR ratings without regressing them to account for error, then you should be aware that there is also the possibility of the same type of error showing up in Clutch ratings.  Taking one at face value can lead to the same problems as taking the other at face value.

"Clutch” is nearly by definition retarded and can be thrown out if it values a Homerun that wins a 1-0 game more valuable in the 8th inning versus the 1st.  It would be more applicable if it was entirely score dependent.

The raison d’etre of clutch is entirely based on the premise that we have no knowledge of the future.

If you decide to then know how a game ended, then clutch doesn’t exist!

So, for people to argue that “in a game that ended...”, and then talk about clutch, well, it doesn’t make any sense.  One precludes the other.

Tango, I can’t see this page quite right.  It is very likely that this is my computer’s fault, because it is very old and needs an upgrade.  Just thought I would mention it in case it wasn’tjust me.  Also, I wanted to mention it because I’m missing the first word or two in all of Kincaid’s sentences.  So basically I’m just giving myself an excuse if I read something wrong/didn’t read something.  Anyway, Kincaid, I agree with pretty much everything you said.  What I was saying was worded poorly.  I was basically saying that clutch had a much smaller measurement error than UZR.  I said it in such a way that sounded like clutch had no measurement error, and that was wrong.  I just a} am trying to think of ways to separate skill and luck without taking away the things that make baseball so damn fun to watch, and b} make my case for Mauer as MVP.  So, in conclusion, I’m better at conceding points than I am at making them.

We were discussing this post on another board, and an important issue was raised that fangraphs’ clutch stat is not a good tool to compare players.  It compares a player’ performance in “clutch” situations to his overall performance.

So a player with a .500 who has a .650 OPS in clutch situations will have a much better “clutch” stat than a player with a 1.050 OPS but a .950 OPS in clutch situations.

For those who actually want to evaluate clutch performance for the purposes of MVP, it doesn’t matter how much better or worse Mauer did than his normal self in these situations.  It matters how he did compared to other players.

I’m doing this quick so I hope I didn’t screw up, but in late inning pressure situations (LIP) and LIP with runners on base (using CBS sportslines’ stats), here’s what the three players have:
Gutierrez in 102 ABs: .919 OPS
Figgins in 106 ABs: 1.096 OPS
Mauer in 81 ABs: 1.163 OPS

It seems clear clutch is a bad stat to use when comparing players.

You are wrong.

Absent the context information, you are giving credit of .13 wins for every Mauer HR.

However, if you include the inning, score, base, out, some Mauer HR will be worth much less than .13 wins, and some will be worth much more.

So, if you START with .13 wins for each Mauer homerun, you then need to add or subtract FROM THAT NUMBER, how much his HR have really been worth.

On the other hand, had you simply started with exactly the win value for each HR as something unique, then you simply get WPA.

Therefore, you have to understand why the numbers work out the way they do:
- WPA gives credit to each HR, on a case-by-case basis
- the context-neutral measure basically gives each HR the same .13 win value
- the difference between the two is called “clutch”; you can call it whatever you want (deviation from typical performance, or quatlus); regardless what you call it, it is a derived value, not calculated

Are you saying I’m wrong in that you CAN use clutch to compare players?  I mean, yes, I guess you can say that you can compare who deviates the most in a positive fashion from typical performance.

But one person could deviate negatively from typical and still outperform someone who’s deviated positively.  It all depends on each of their own individual baselines, correct?

So one could say that Gutierrez has demonstrated more clutchness via the fangraphs stat, but you’d still rather have Mauer up to bat in those situations correct?

If not, what am I misunderstanding?

Yes, this is no different than saying that Scutaro or Crede is more clutch than ARod, but you’d still choose ARod with the game on the line.

That however is irrelevant to the point you originally made.

Thanks for responding so quickly.  I’m still not quite following.  Reading your post, I gathered you were saying, “since Gutierrez and Figgins are more clutch than Mauer, that is points in there favor for MVP voting.”

I believe that the clutch stat was developed to help aid in determining whether clutch was actually a skill or just a lucky circumstance that fluctuates often with all players, no?

But that seems to me to be clouding the issue of MVP discussion.  Again, apologies if I’m missing something, but as I see it, when thinking about clutch for MVP, I would think to myself, “who has performed best in high leverage situations?” I wouldn’t think to myself, “who has most improved their performance over their typical performance in high leverage situations?”

So my point was, regardless of clutch, I’ll take Mauer with his 1.163 OPS in those situations over the other two, regardless of if his OPS is higher in other situations.

And I know the 1.163 is not necessarily high leverage, as it’s just CBS’ LIP point of view, whatever that is.  But I guess regardless of what Mauer’s actual OPS is in high leverage situations (or whatever other stat one prefers, OPS+, WOBA, etc.), if it’s better than the other two, wouldn’t you rank him higher when discussing clutch for MVP purposes?

How can I say this?

Mauer, if you forget about the timing of his hitting, is roughly +5.0 wins above average as a hitter.  If you include when he hits what he does, he’s roughly +3.6 wins.

Gutierrez, if you forget about the timing of his hitting, is roughly +0.6 wins above average.  If you include when he hits what he does, he roughly +3.4 wins.

So, component-wise, you’d state it like this:
Mauer:
- context-neutral is +5.0
- clutch is -1.4
- TOTAL (which is all we care about) +3.6

Gutierrez:
- +0.6
- +2.8
- TOTAL +3.4

If you ONLY consider teh context-neutral, then it’s Mauer at +5.0 and Gutierrez at +0.6.  If you ALSO include Clutch, you make an ADJUSTMENT of -1.4 for Maur and +2.8 for Gutierrez.

Mauer still is ahead, as a hitter.

Gutierrez demolishes him as a fielder.

But the thing that’s not jibing in my head is what that -1.4 means and what the +2.8 means.

How can Mauer have a better OPS, and it’s MUCH BETTER OPS, in high leverage situations than Gutierrez and yet Gutierrez is raised to within .2 wins of Mauer specifically FROM those high leverage situations?

Because not all hi-lev situations are the same.

I’m not following here, either.

wRAA: 52.7
RE24: 52.58
WPA/LI*10: 50.5
WPA/pLI*10: 38.6

So it seems that all this presumes that the “correct” value for a hitter is WPA/pLI.

Now - and you can feel free to step me through where I’m wrong here, but looking at his situational stats:

http://www.baseball-reference.com/players/split.cgi?n1=mauerjo01&year=2009&t=b

It really feels to me like Mauer’s “Clutch” score isn’t a reflection of his poor performance in high LI spots, but the simple fact that he has so few high LI spots relative to his low LI spots.

I will explain this poorly, for the sake of explaining it at all. I will come back for it later to clean up my mess.

Gutierrez has a 3.30 WPA/pLI. Mauer has a 3.86 WPA/LI. Gutierrez has 383 outs, compared to Mauer’s 263 outs. (I figured outs as (1-OBP)*PA, which isn’t quite right but close enouth to illustrate.)

Now of course if we’re plugging WPA/pLI into WAR in place of wRAA/Runs_Per_Win then hypothetically we’re crediting them for the difference in the number of outs they’re creating. Right?

The problem is that we’re crediting batters for their follow-on PAs given average leverage, which is simply not true - we know that Mauer is giving his follow-on hitters a higher LI than Gutierez is.

So in order to accurately leverage adjust, we need to give both hitters credit for the average LI of their follow-on hitters, after adjusting for the average LI of each hitter’s PA.

Could this explain it?  Is mauer actually getting penalized with negative clutch for performing well in low leverage situations?

http://www.fangraphs.com/forums/showthread.php?t=2962

if that’s the case, then my original statement would be correct that clutch may be useful for determining if a particular player has an actual skill called clutch but it’s only relevant as a comparison with himself, not with others.

The poster on fangraphs might be completely wrong, and I confess I’m not experienced with the underlying stats, but on the face of it this is just not making sense to me given the huge difference in their actual performance.

A diffference of that many wins and I would expect gutierrez to be all over the papers for his clutchness and mauer to be all over th papers for his suckitude in clutch situations a la robinson cano, who in fact does have a negative clutch and who’s poor clutch performance this year is constantly discussed during games.

But gutierrez isn’t even in the top 20 in BA in “close and late” situations.

Color me confused.

Yes, performing better in lower leverage situations is more or less what a poor clutch rating is.

If you look at Mauer’s context neutral production, it’s great.  It’s much better than Franklyn Gutierrez’ context neutral production.  If you want to include the leverage index of the situations, then Mauer’s production will go down because all that awesome production he put up in low leverage situations is worth less once you look at leverage index.  Mauer’s negative rating is how much you subtract from his context neutral total because it over-counts his awesome low-leverage production.  Gutierrez’ positive rating is what you would add to his context neutral total.  Mauer’s starting point is way higher than Gutierrez, though, so he can still rate as a better hitter even after including clutch.

All you are doing with clutch is seeing how including leverage index affects the production for that individual hitter, and how much it raises or lowers his context neutral total.  It has nothing to do with what his starting point of context neutral production is.

As I said, if people have a problem with calling the calculation clutch, then call it “deviation from typical performance” (DFTP).

***

“But gutierrez isn’t even in the top 20 in BA in “close and late” situations. “

I don’t know if you got the memo, but if you quote batting average around here, you are banished for an eternity.

As I said, not all close/late situations are equally created.  If Ichiro had four of those walk-off HR, he gets +3.44 in WPA, when he would only otherwise get around +0.52 wins if they were random.  So, his DFTP would come in at close to 3 wins, baed on those 4 PA.  This means that he can be his normal self in all the other hi-LI situations, have OBP/SLG that look “normal” in hi-LI situations, and STILL get a DFTP of +3.

It is what it is.

Here’s Eric Seidman explaining clutch:
http://www.fangraphs.com/blogs/index.php/all-about-clutch

So you’re subtracting context neutral performance from high leverage performance.  Why do this?  In order to find out if that player does better in those situations.  Not better than OTHER PLAYERS.  Better than HIMSELF.

So the assertions in the original post, “He has not performed in high leverage situations very well (relatively speaking)” and “if you bring clutch into the discussion, all of a sudden, you are introducing Franklin Gutierrez, who has been sensational in the clutch” are incorrect.

Gutierrez has NOT been sensational in the clutch.  He has been an exact league average hitter all year with exactly a 100 OPS+ according to baseball-reference.

In the clutch, perhaps he’s been a 120 OPS+ player.  I’m not sure, but it’s certainly not 180.

Mauer, OTOH, has had a 182 OPS+ for the year, and perhaps his OPS+ in these situations is somewhere near 160.  Whatever the actual number is, he is blowing Gutierrez out of the water in these situations.  He just happens to be underperforming Everyday Joe Mauer.

You are giving Gutierrez credit for being league average at all times and penalizing Mauer for being incredible.

Mauer has performed like an all star in the clutch and Gutierrez has performed as an above average player in the clutch.  Context neutral, Mauer has performed as an inner circle hall of famer, Gutierrez has performed like a player who might not get another contract (not including defense of course.)

Just look at baseball reference’s situational hitting:
Mauer:  http://www.baseball-reference.com/players/m/mauerjo01-bat.shtml#batting_situational

Gutierrez:  http://www.baseball-reference.com/players/g/gutiefr01-bat.shtml#batting_situational

Mauer has scored 20% of his baserunners vs. league average 14%.  Gutierrez has scored a whopping 15% of baserunners.  Even more of a gap when you look at the stats in particular situations.  Mauer’s scored 66% of runners on 3rd base with less than 2 outs.  That’s awesome.

It seems to me the ranking of their performances would be:
Everyday Joe Mauer
Clutch Joe Mauer
[insert wide gulf here]
Clutch Franklin Gutierrez
Everyday Franklin Gutierrez

Those original assertions should be, “He [Mauer] has not performed in high leverage situations very well when compared to his normal ridiculously stellar performance,” and “if you bring clutch into the discussion, all of a sudden, you are introducing Franklin Gutierrez, who has been sensational in the clutch when compared to his normally pedestrian performance.”

"I don’t know if you got the memo, but if you quote batting average around here, you are banished for an eternity.”

LOL.  Yeah, it’s the only stat I could find and even number 20 is at .330.  http://stats.oregonlive.com/mlb/getleaders.asp?rank=056

I was just alluding to the fact that if Gutierrez should be getting this much credit, he should have eye popping stats all over the place.

No, he doesn’t have to have it “all over the place”.  He simply has to have it more than his overall stats would suggest.

He’s been terrible in low-leverage situations (when we dock him more than we should) and great in hi-lev situations (when we don’t credit him enough).  The combination of the two is enough to show one of the highest deviations from typical performance (DFTP) that you will ever see.

(And in no way do I believe that this is replicable.  It happened, like winning a lottery.)

This is alot like arguing between:
- richest man in the country
- most-skilled investor in the country

Which is the “most valuable” in money-making?

"The combination of the two is enough to show one of the highest deviations from typical performance (DFTP) that you will ever see.”

Yes, but you’re still not responding to my main questions:
1. Did Mauer outperform Gutierrez in those situations, i.e. is my ranking correct?
Everyday Mauer
Clutch Mauer
[gulf]
Clutch Gutierrez
Everyday Gutierrez

2. If so, why would Gutierrez’s clutch stat have anything to do with an MVP discussion?  For it to be relevant to an MVP discussion, Clutch Gutierrez would have to outperform Clutch Mauer.

I was still wondering if this was correct as well.  “To use clutch in the definition of “valuable” implies that you actually do believe that clutch hitting exists. Because the situation the individual performer came up to bat or pitch in was not in their control. You can’t give them credit for being allowed to bat in high clutch situations. It is almost completely a product of other people’s performances that produces clutch situations.” The MVP award is not, who is the most valuable player to their team, its who is the most valuable player to any team.  IMO

According to B-R, Gutierrez has outperformed Mauer in clutch (high lev) situations:

Gutierrez (105 PA): .349/.446/.581/1.027, 173 tOPS+ (compared to himself in all PA), 174 sOPS+ (compared to league in high lev)

http://www.baseball-reference.com/players/split.cgi?n1=gutiefr01&year=2009&t=b#lever

Mauer (85 PA): .333/.424/.514/.937, 80 tOPS+, 151 sOPS+

http://www.baseball-reference.com/players/split.cgi?n1=mauerjo01&year=2009&t=b#lever

"Yes, but you’re still not responding to my main questions: “

Patrick/52 has the correct data, so I don’t know why you drew the list you did.

"In the clutch, perhaps he’s been a 120 OPS+ player.  I’m not sure, but it’s certainly not 180. “

174 close enough?

As I said, not all close/late situations are equally created.  If Ichiro had four of those walk-off HR, he gets +3.44 in WPA, when he would only otherwise get around +0.52 wins if they were random.  So, his DFTP would come in at close to 3 wins, baed on those 4 PA.  This means that he can be his normal self in all the other hi-LI situations, have OBP/SLG that look “normal” in hi-LI situations, and STILL get a DFTP of +3.

Okay, but doesn’t WPA/LI account for that? In fact, doesn’t WPA/LI account for that better than WPA/pLI, which has NO idea what the LI is for any of those individual PAs?

Now - and you can feel free to step me through where I’m wrong here, but looking at his situational stats:

...

It really feels to me like Mauer’s “Clutch” score isn’t a reflection of his poor performance in high LI spots, but the simple fact that he has so few high LI spots relative to his low LI spots.

The number of high LI or low LI opps has no impact on the overall sum[ WPA(i)] / LI(i) ] total, where i=a particular PA.  By dividing by LI, we are rescaling the WPA figure such that it makes each PA equal in impact.

Ah, perfect, Patrick.  Thanks!

I didn’t see the leverage section.  That split page rocks!

So in high leverage situations, Gutierrez wins.  In medium, Mauer wins by a wide stretch.  And in low, it’s simply no contest; Gutierrez is a freakin’ liability.

My rankings may be wrong, but I guess it depends on where you put that “medium leverage” section.

But my main point still stands.  Mauer has an OPS+ of 151 in high leverage situations, and 174 in medium leverage situations.  So the statement “He has not performed in high leverage situations very well” is not correct.  He has performed very well.

I do stand corrected on Gutierrez in that he’s been ridiculously good in high leverage situations.  But tit’s clear that it’s not a valid use of clutch to add or subtract it from wins above average to compare players.  Mauer is still far above average in clutch situations.  151 OPS+ is clear hall of fame caliber over a career, and it’s clearly above average in any situation.

sum(WPA)/pLI is used to establish the baseline of all LI = 1.

sum(WPA/LI) is used to establish the baseline of EACH LI = 1.

The difference is the DFTP.

Think of a reliever, who has a WPA of +4.5, LI of 1.5, and WPA/LI of 3.0.

If I don’t do 4.5/1.5 first, he’s going to look like he has a DFTP of 1.5 (4.5 - 3).

The number of high LI or low LI opps has no impact on the overall sum[ WPA(i)] / LI(i) ] total, where i=a particular PA.  By dividing by LI, we are rescaling the WPA figure such that it makes each PA equal in impact.

Yes, but when we compute WPA/pLI what we are doing is:

[ WPA(i) / AVG(LI) ]

instead, right?

So the number of high/low LI opps does NOT affect WPA/LI, but it SHOULD affect WPA/pLI. So when we take the difference between WPA/LI and WPA/pLI (whether you call it “Clutch” or “DFTP” or “quatloos") aren’t we really taking the difference of the players’ OPPORTUNITIES, not their PERFORMANCE?

No/57:

In total, if you were to weight their performance by how much win impact they had, at the moment they performed, Gutierrez and Mauer are equivalent.

Gutierrez poor performance in low-leverage situation simply has very limited impact.

Basically, it’s like counting his hi-LI performance as “3” for each PA, and his low-LI performance as “1/3” for each PA.  Something like that.

And if you look at each individual PA one at a time, counting them by exactly their LI, then Gut and Mauer are equivalent.

PLEASE, read the link in Post 1, as Mike Emeigh gets how this works:
http://www.insidethebook.com/ee/index.php/site/comments/leverage_index/#1

"Gutierrez poor performance in low-leverage situation simply has very limited impact.”

But Mauer’s high performance in low leverage has a high negative impact.  That makes zero sense.

However you feel about the rest, Mauer’s OPS+ in high leverage situations is far above average.  Your original statement is wrong.  If Gutierrez has a 140 OPS+ in high leverage, he would still have a very positive clutch number and the way you used it to make your argument would still hold.  But it would be obviously wrong.  So whether it’s 178 or 140, Mauer has hit very well in the clutch, and so has Gutierrez.  The difference is that Mauer has been great in every situation, and Gutierrez has been average or terrible in some.

Put another way, you can make Mauer’s clutch number go up in one of two ways: Improve his OPS+ in high leverage situations to over 191, or reduce his medium leverage and low leverage OPS+ to under 151.  Both have the same impact on the clutch number, and yet each means very different things in terms of the value of Mauer.

Colin/59: how do you interpret Tango/58?

I said that the two players have a similar hitting value, if you take context into account (aka WPA).

I also said that you can break this number down into two components:
1. the traditional way of looking at their hitting numbers
2. the difference between the traditional way and WPA

The question seems to be how to interpret that #2, calling it clutch, DFTP, quatloos, or any word in the universe you want to use.

But, the ONLY thing that we are interested in is whether to use WPA or #1 above.  Using WPA automatically uses #1 and #2.

So, do you want to use:
- #1 only
- #1 and #2

OK, I think I’m getting to the bottom of this.  I’m new to these stats as I said, so please correct me if I’m wrong on any of this.  I do think I’m generally not dumb when it comes to stats, so the conversation might be helpful to other newbies, as if I’m misunderstanding something I’m probably not the only one.

If we look at the components of clutch, let’s start with WPA.  Doesn’t WPA already have context built in?  So Gutierrez is close to Mauer in WPA for a number of reasons:
-he’s got a ridiculous 178 OPS+ in high leverage situations
-Mauer gets not much credit for his 191 OPS+ in low leverage situations
=Gutierrez does not get penalized very much for his terrible 57 OPS+ in low leverage situations.

So Gutierrez is already being brought way up to the front of the class because of these distortions in WPA.

Tango, you even explain this pretty fully in the comments on this thread:
http://www.baseballthinkfactory.org/files/newsstand/discussion/kapur_on_wpa_or_why_johnny_damon_is_your_2009_al_mvp/

where you state “WPA tells you the events that caused the biggest cheers (or moans). It goes farther in that it also tells you WHO was involved. And it goes farther still by TRACKING who was involved over the season. And it goes even farther by compiling leaderboards.”

So in essence, I think you would agree that Gutierrez has no business being on anyone’s leaderboards for overall offense, but one can still marvel at the fact that “hey, look at this.  Franklin Gutierrez has happened to be involved in a lot of game changing plays this year, and he’s done pretty damn well in those situations.”

And yet, Mauer’s WPA is still higher than Gutierrez.  I might stop here actually because for actual clutch comparison, shouldn’t you just use WPA? 

Now isn’t dividing WPA by LI meant to remove context?  Or place a player’s performance into the “average context seen by the league?”

So I believe WPA measures context included performance
WPA/LI measures context neutral performance
clutch measures the difference between the two.

So looking at clutch helps answer the head scratching you get when you look at the WPA leaderboard (it answers the question, “How in the heck can Franklin Gutierrez be in the top 10 in WPA?”, but the “clutchness” is already integrated into WPA, so why go any further?

Mauer is already “undercredited” in WPA for his performance in low leverage situations and Gutierrez is already “overcredited” in WPA for his performance in high level situations.  Further, Mauer is penalized in WPA vis a vis Gutierrez by seeing a lower percentage of high leverage ABs (15.6% vs. 18.2%).  Plus Mauer’s missing a month reduces his WPA.

By isolating clutch and calling it out from WPA for comparison purposes, you are in a way doubly crediting Gutierrez and doubly penalizing Mauer, no?

Tango/63: Ah, well I guess you said the same thing as I said in 64, just much shorter.  LOL.

But don’t you think then that my last question is valid?  “By isolating clutch and calling it out from WPA for comparison purposes, you are in a way doubly crediting Gutierrez and doubly penalizing Mauer”?

Clutch helps explain why Gutierrez is high in WPA, but it doesn’t make him anywhere near an MVP, and it shouldn’t be used when evaluating MVPs.  If one cares about clutchness/context, one should stick with WPA.  But no one in their right mind should use WPA as the sole measure of MVP!

Clutch helps explain why Gutierrez is high in WPA, but it doesn’t make him anywhere near an MVP, and it shouldn’t be used when evaluating MVPs.  If one cares about clutchness/context, one should stick with WPA.  But no one in their right mind should use WPA as the sole measure of MVP!

Anyone who uses WPA as the “sole” measure is being silly, since you are not including fielding, position, or replacement level.  Your statement there is a red herring.

Now, back to my question here:

So, do you want to use:
- #1 only
- #1 and #2

If you have ALREADY decided that you don’t care about the timing of the offensive performance, then you have already precluded WPA (and by extension clutch) into the conversation.

If you have ALREADY decided that you care deeply about Gutierrez being lights out with the game on the line, then you care only about WPA (and by extension clutch).

So, the discussion of clutch is really irrelevant, and a byproduct of your decision: do you care about context-neutral, or context-specific performance (with respect to the context of inning, score, base, outs)?

There is no wrong answer here.  All that happens is that for those who care about context, they cannot, by definition, have a conversation with those who don’t care about context.  It will be the very definition of talking past one another.

OK, so I think we’re on the same page.  Now let’s go back to your original statements:
“He has not performed in high leverage situations very well (relatively speaking).  His performance, as eye-opening as it looks, if you look at WHEN he performed, has less oomph…
Source: Fangraph’s Clutch, which I helped define.  (I think I even invented it, but maybe one of my readers did, and I modified it?  I don’t remember any more.”

“However, if you bring clutch into the discussion, all of a sudden, you are introducing Franklin Gutierrez, who has been sensational in the clutch.  Or, a good clutch season by Chone Figgins.”

These two statements make me wonder if you sorted AL leaders in clutch and chose Gutierrez and Figgins as examples that must be considered if clutchness is important to you for MVP.

I’m arguing that this is not true.  The clutch stat doesn’t matter, or, it doesn’t matter unless you are looking for reasons to explain why a particular player has a lower or higher than expected WPA.

If you want to make a list of clutch leaders as one of your attributes for evaluating MVP qualifications, you should use WPA.  And if you DID use WPA, you would say that Mauer is a better candidate than either of the guys you mentioned in the clutch department.

True, or am I off somewhere?

And if you DID use WPA…

I did (so you can ignore the rest of your post leading to that).

..., you would say that Mauer is a better candidate than either of the guys you mentioned in the clutch department.

True, or am I off somewhere?

You are completely ignoring position and fielding and playing time.

I’m not talking about “Most Outstanding Hitter” award.  It’s PLAYER.

Tango/68: You ignored fielding as well.  I’m dealing with your original post.  I’m not making a vote for MVP in my comments.  Yes, I would vote for Mauer, but I haven’t raised the issue of my own opinion in my posts.  I’ve just been trying to understand why you would use clutch, and now you’re claiming you’re not.

When you brought up Gutierrez, you brought him up specifically for his clutchness, not for his fielding.

And when you say “I did [use WPA] (so you can ignore the rest of your post leading to that)” then you must at least concede that by putting “Source: Fangraph’s Clutch”, you caused a lot of confusion for your readers.  Why not say, “Source: WPA”?  Or why not just say, “Crap.  I used clutch when I shouldn’t have.  Sorry folks.”

I have to say I have a real tough time believing you used WPA since you never mentioned it.  If you used WPA, you wouldn’t negatively compare Mauer to Gutierrez or Figgins, since he’s above them both and Figgins is way down at 13.  [No disrespect meant, and I do really appreciate you continuing this dialogue.  It just doesn’t seem to jibe with your original post.]

And you still haven’t responded to my statement that “He has not performed in high leverage situations very well” is at best an overreach and at worst simply not true considering he has a 151 OPS+ in those at bats.  I have a hard time believing you would look at a player’s OPS+ and characterize 151 as not performing very well.  If you fall back to the “not all high leverage ABs are created equal” argument you would end up reducing the entire “Mauer has not performed well in the clutch” argument to probably less than 10 very high leverage at bats where he made an out.  I mean, he’s got only 85 PAs as it is with leverage over 1.5.  And could anyone ever really have an MVP discussion based on less than a week’s worth of ABs?

Just look at my post number 33, where I still wasn’t quite sure of myself.  I ended with “It seems clear clutch is a bad stat to use when comparing players.” You began post 34 with simply “You are wrong.” I assumed that to mean “you are wrong about clutch”, which seems the obvious conclusion.  And that’s what I’ve been debating ever since.

Tango/68: You ignored fielding as well.  ... I’ve just been trying to understand why you would use clutch, and now you’re claiming you’re not.
When you brought up Gutierrez, you brought him up specifically for his clutchness, not for his fielding.

I never ignore fielding.  Or position.  Ever.  You are obviously new around here.

I brought up Gutierrez because his WAR (which is his neutral-hitting, fielding, position, replacement-level) plus his Clutch is so high.

Alternatively, I could have used WPA in place of his neutral-hitting plus Clutch to get the same answer.

And when you say “I did [use WPA] (so you can ignore the rest of your post leading to that)” then you must at least concede that by putting “Source: Fangraph’s Clutch”, you caused a lot of confusion for your readers.  Why not say, “Source: WPA”?  Or why not just say, “Crap.  I used clutch when I shouldn’t have.  Sorry folks.”

Components, components, components.  Because his context-neutral hitting PLUS clutch IS WPA.  And in my post, I said this:

His performance, as eye-opening as it looks, if you look at WHEN he performed, has less oomph.  Enough to knock out 1 win from what the performance otherwise looks like.

Source: Fangraph’s Clutch,

So, the “knock out 1 win” refers to Clutch.

Yes, I may be confusing readers, but that’ll have to be the case.  My usual readers know pretty much what I mean, and I know that Neyer sends alot of readers this way, but I only know this after the fact.  A new reader may get confused.  That’s part of the growing pains.  You’ll be rewarded if you stick around.

As Bill James once noted, if it seems like you are walking in the middle of a conversation, you are.

“He has not performed in high leverage situations very well”

Relatively speaking.  Relatively speaking.  Hang me that in the 1000 words I write on my blog every day with no editing or proofreading, I forget two.  But, OBVIOUSLY he’s a great hitter no matter what.  And just one look at his stats in high-lev situations shows he’s an above average hitter there.  So, if you can grant me the benefit of the doubt that my statement can be made true by simply adding “relatively speaking”, that’d be great.

Otherwise, I’m going to have to qualify everything I write to make sure it’s all unambiguous.

Everything, in any case, is relatively speaking to something.  I simply didn’t say relative to himself, which is typical in clutch discussions.

So, and now you’ll have to forgive my non-understanding of these stats if they are indeed counting this, but why isn’t Mauer’s low leverage hitting counted as much?  If he drives in three runs in the first and it turns out later that doing so created a 3-3 tie in the 9th so that Nick Punto could deliver in the clutch who gets more credit there?

Kyle:

Please follow the thread here:
http://www.insidethebook.com/ee/index.php/site/comments/wpa_captures_the_essence_of_baseball/

If your comment still applies, please ask it there.

I’ll have to quibble with two things;
“I never ignore fielding.  Or position.  Ever.  You are obviously new around here.”

I know who you are, believe me, which is why I’ve been attempting to be respectful, as I’m pretty psyched you’re entertaining a debate with me.

But I’m lefty, so I’m going to keep pushing.  I can’t help myself.

“Everything, in any case, is relatively speaking to something.  I simply didn’t say relative to himself, which is typical in clutch discussions.”

And there is the rub.  Why would “relative to himself” have anything to do with an MVP discussion, even if you’re discussing clutch?  It wouldn’t.  That’s been my point all along.  That couldn’t have been what you meant because someone in your position and knowledge could not believe that “relative to himself” would be relevant in an MVP discussion.  Relative to everyone else is what matters.

I said it above, but again, I might say to myself, “What the heck?  Mauer’s stats are ridiculous.  Why is his WPA so low?  [research, research, research.] Ah. Negative clutch.” But that’s a whole lot different than saying, “Mauer has not performed very well in high leverage situations.”

"Why is his WPA so low? [research, research, research.] Ah. Negative clutch.”

No, I think that’s incorrect. The reason why Mauer’s WPA is “so low” is because of opportunities (WPA is a counting stat, obviously). Mauer missed most of April, so between that and random chance, he only has 85 high leverage PA this year (just focusing on high-LI for simplicity).

While he’s done well in those situations (151 OPS+), some have had many more chances (Andre Ethier has 149 high-LI PAs, and has taken advantage to the tune of 132 OPS+, including 6 walkoff hits, 4 of them HRs)

Others have had more high-LI PAs AND better results in those PAs: Chase Utley has 103 PAs, and blows Mauer away with a 197 OPS+ in those PAs. Ryan Howard: 210 OPS+ in 127 high-LI PAs.

Any clutch-related narrowing of the gap between Mauer and the rest of the AL position player field (such as Gutierrez or Teix or whomever), is not because Mauer has been unclutch relative to himself, but rather that others have provided more clutch value to their teams through some combination of performance and opportunity.

If Mauer performed equally well in clutch and non-clutch situations, given his opportunities, he’d be around +5.0 wins.

Given his actual performance (WPA) he’s at +3.5 wins.

That difference, using whatever word in the universe you want, is -1.5 wins.

Anyone who wants to see how Gutierrez or Mauer got the WPA ratings they have can go to Fangraphs and look for their playlogs.  It will tell you what the details of each game situation they faced were, and the resulting WPA.

There is one play, high on Gutierrez’s list, where he got a lot of positive WPA for hitting into a bases loaded error in a tight late inning spot.  That’s a little questionable, but the same rules apply to everyone, every error in such a situation will add to player WPA, and it’s an isolated case, his other high WPA plays were big hits, often homers, in big situations.

I remain am a firm believer in using WPA as a way of measuring value (along with position, fielding, replacement etc.)Not everyone’s opinion, just mine. Here is why:

Definition (found on line)

val·u·a·ble (vly--bl, vly-)
adj.
1. Having considerable monetary or material value for use or exchange: a valuable diamond.
2. Of great importance, use, or service: valuable information; valuable advice.
3. Having admirable or esteemed qualities or characteristics: a valuable friend

To me #2 sticks out “IMPORTANCE”.

To me WPA measures IMPORTANCE of an event (ie: hit / walk / homer), obvious elevated by the situation / context, which I think we all agree is out of the hands of the individual batter. But, as stated already this is not an award for Best Player - as that would be simply mailed to Pujols in April.

Sometimes the most valuable advice is not the ‘best’ advice, it’s the most ‘important’ advice. I could spend a fortune talking to the most skilled financial planner in the country about where to invest my money, but I could also just listen to my 85 year old grandfather who told me to hide it under my mattress. If I took the latter advice in July 2008 and looked at my statement in March 2009, that would have been the most valuable advice I could have gotten. Neither could have predicted the Lehman Brother collapse in September (context), and 9 times out of 10 the financial planner would be correct. However, over the period of July 2008 to March 2009 the adivce from my grandfather was the most valuable to me as I ended up with the most money (wins), regardless of his skill at investing.

Patrick/74: “‘Why is his WPA so low? [research, research, research.] Ah. Negative clutch.’

No, I think that’s incorrect. The reason why Mauer’s WPA is “so low” is because of opportunities (WPA is a counting stat, obviously). Mauer missed most of April, so between that and random chance, he only has 85 high leverage PA this year (just focusing on high-LI for simplicity).”

I was also just focusing on clutch for simplicity.  There are three reasons: playing time, an unusually low number of high leverage situations (25% less high leverage ABs than typical for someone with his amount of ABs) and clutch.  And by clutch I mean clutch illustrates that mauer has performed better in low leverage situations than high leverage situations, reducing the WPA he would have if he performed equally well in all situations.  If he had a clutch of zero he would still be reduced by the other two reasons of course.

Question RE pitchers and WPA.  I believe Fangraphs still simply credits/debits all defensive WPA to the pitcher due to limitations in the play by play data.  Is there any idea if this tends to over inflate WPA for starters, underinflate, or balance out over a season?

Second question: Imagine a player hits a leadoff home run in the top of the first, and the team eventually wins 1-0.  Why can’t that player continue to get additional tiny slivers of WPA as the game wheres on?  As the game gets later and later, that value of that homer increases, but the people who begin to get credit are the pitchers.  Indeed they should get a bunch of credit for continuing to maintain the shutout, but I imagine the WPA added per out/inning is higher in a 1-0 game than it is in a 0-0 game.  Shouldn’t that tiny difference in the WPAs between 1-0 and 0-0 go to the guy who hit the homer, not the pitcher, as he’s a very big reason for the continued increase in win probability.  And since you can simply subtract the win probably of each out if the score were 0-0 from the win probably of each out when it’s 1-0, it seem easy to calculate the hitter’s contribution.

Perhaps it just gets far too complicated as more runs are added to both teams.

I don’t know if I’m explaining my thinking correctly, and perhaps there’s a better thread for this, but this occurred to me, so I posted.

See link in post 72.  If your question still applies, please post it there.